how to get a bash shell with the new install floppies (Re: new boot floppies being uploaded)
I'd submit the following as a bug report, but i don't know what
package/version to assign it to.
IMO that should be documented on the installation disks, a notice saying
"if you find any bugs then send mail to submit@bugs.debian.org with the
following details:....."
On Sun, 8 Dec 1996, Bruce Perens wrote:
> I am uploading a new boot floppy set 1996-12-8. Fixed are:
>
> The inittab bug. The sysvinit package changed and I didn't notice.
> modconf complaints about tset, "*" in ipv4 menu, bad drawing in fs menu.
> /dev/MAKEDEV complained about its configuration file.
> /root/.bashrc missing.
this fixes most of the problems i had with the 1996-12-7 set. I'm still
having a few problems with the new disks:
1. ftp is still missing from the base install set. This makes it
difficult to install a custom compiled kernel. Actually impossible
if you don't have your kernel available on an nfs mount or floppy
disk.
Also, there are some circumstances where it is necessary or
convenient to install a custom package before running dselect (e.g.
a custom-built package to preconfigure the defaults & config files
which subsequent packages expect). ftp would make that a lot easier.
2. MBR doesn't work for me. No matter what I do, I can not get a machine
to boot linux from the hard disk if MBR is in the master boot
record.
This is consistent - it's happened on at least the last 10 or so
debian machines i've built (mostly AMD 586-133 pci boxes with IDE
drives, but also some 486 & pentium machines...some with aha1540 or
aha2940 or ncr53c810 cards and scsi drives).
Note, this is not new with debian 1.2 - it also happened with 1.1 &
unstable over the last few months.
Fortunately, I've figured out a workaround which may be useful to
others:
After the base disks have been installed onto /target, do the
following:
- press Alt-F2 and press Return to get a sh prompt on tty2.
- type:
cd /target
export LD_LIBRARY_PATH="/target/lib:/target/usr/lib"
./usr/sbin/chroot /target
this gets you a bash shell chrooted to /target. The environment is
very similar to what you'd get after booting - except that NONE of
the /etc/init.d/* scripts have been run. If you need networking at
this point, for example, you'll have to run /etc/init.d/network
yourself.
The thing to remember is that you're working in a semi-configured
environment. Don't expect things to work exactly as they would
in normal circumstances. This is really only useful for whatever
emergency hacks/fixes you need to get the system to boot.
- From here you can use ae to edit /etc/lilo.conf to change the line
boot=/dev/hda1
to
boot=/dev/hda
- run "lilo" to install lilo as the master boot record.
Note, this is useful for more than just running LILO. You can use
this bash chrooted environment to do pretty much anything you need
to do before rebooting...including installing a new kernel_image if
you have one on a floppy disk or nfs mount. If you need more virtual
ttys you can open as many as you need with the "open" command.
It would be nice if this was either documented with the disk set or
if a sh script was included which made it easy - possibly as a menu
option to invoke the chrooted bash shell on tty4 or tty5.
craig the nitpicker (who is otherwise very happy with & impressed by the
new debian 1.2 install procedure...a great job, overall!)
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-REQUEST@lists.debian.org . Trouble? e-mail to Bruce@Pixar.com
Reply to: